INDEX

Explanations

toll gates and boarding gates

The neuron fires on occurrences of the word “toll” (including in compounds like “toll‐like receptor”) and related payment‐collection terms.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ANSWER

-0.78

 víz

-0.72

Carro

-0.66

 Unsafe

-0.66

 Cream

-0.66

 obscurity

-0.65

 Bambi

-0.65

SKY

-0.64

 yıldız

-0.64

 låter

-0.64

POSITIVE LOGITS

 gate

1.69

gate

1.59

gates

1.52

 gates

1.45

 Gate

1.35

Gate

1.23

 booths

1.16

house

1.07

booth

1.06

Gates

1.05

Activations Density 0.042%